Scholarly Reading by Faculty in the United States: Summary Results of a Study Conducted in 2012 in Five Universities

Transcription

1 Scholarly Reading by Faculty in the United States: Summary Results of a Study Conducted in 2012 in Five Universities Carol Tenopir, Rachel Volentine, and Lisa Christian Center for Information and Communication Studies University of Tennessee, Knoxville Spring 2013 Funding by the Institute of Museum and Library Services (IMLS)

2 Contents Executive Summary and Key Findings 3 Introduction 5 Previous Studies... 7 Methodology... 9 Demographics of Respondents.. 12 Scholarly Journal Article Reading 23 Difference of Article Reading Patterns by Demographics 44 Scholarly Book Reading 62 Difference of Book Reading Patterns by Demographics.. 74 Other Scholarly Publication Reading Difference of Other Publication Reading Patterns by Demographics.. 94 Social Media: Participation and Creation. 105 Open Ended Questions 119 Role of Library Collections 149 Bibliography Copy of Survey

3 Executive Summary The Lib-Value project measures the value, outcomes, and return on investment of academic library collections and services. This report measures the value of the library collections by examining the scholarly reading patterns of faculty members in the United States and comparing their use of the library with other sources for scholarly materials. Starting in February 2012 through October 2012, faculty members at five universities in the United States were invited to participate in a survey of their scholarly reading behavior. We received 837 responses from a total faculty population of 11,332 for an overall response rate of 7.4%. Any conclusions must be made cautiously due to this low response rate. The survey asked questions about reading of articles, books, and other scholarly materials from all sources (library-provided, other sources, and social media), and focused on use value (outcomes of reading) and exchange value (time spent obtaining and reading). Important findings include: Over half (54%) of article readings by United States faculty respondents are obtained from a library or school/department subscription, and 93% of those obtained through a library or a school or department collection are from electronic collections. While faculty members prefer electronic resources to obtain information, print is still a popular means for reading. Just over half (51%) of article readings are read on-screen, while nearly as many (48%) are read on printon-paper. Just 8% of book readings are ebooks. Faculty who received awards and published more items in the last two years read more articles, books, and other publications. Faculty purchase books (39%) more often than they obtain them through the library (22%). The majority of article readings (52%) are for the principal purpose of research, while book readings are for research (41%) and teaching (34%); other publications (such as reports, government documents, magazine 3

4 articles, or conference proceedings) are read more for current awareness (34%) and research (29%). Faculty members participate in social media more than they create it; however, their use and creation is more often occasional rather than on a regular basis. Faculty who participate or create content for social media tools are reading more articles, books, and other publications. Social media has not replaced traditional articles and books, although faculty members recognize their value in inspiring new ideas. United States faculty members on average spend 121 hours per year of their work time with library-provided material, or the equivalent of 15 eight-hour days annually. 4

5 Introduction The project in context: previous studies and methodology 5

6 In an age of continually growing digitization, globalization, and abundant information, the value of scholarly information remains high to support the work of faculty members. Scholarly material adds value to the quality of their work and guides their future research. Academics now have many choices of where and how to access scholarly articles, books, or other materials. Time, cost, and electronic availability are all factors in their decisions of which materials to select, and by providing the highest-quality material in a convenient manner, the library can ensure they are receiving the best material. This study seeks to answer broad questions such as: Why do faculty members read scholarly materials and do reading patterns vary according to purpose of reading, source of reading, or individual characteristics of readers such as academic discipline, status, or age? What is the role and value of the college and university library in providing access to scholarly content in this changing digital landscape? The Value, Outcome, and Return on Investment of Academic Libraries project (Lib- Value) is a three-year study funded by the Institute of Museum and Library Services (IMLS). Part of the project seeks to measure the value of the library s provision of access to scholarly materials by examining scholarly reading patterns and comparing use patterns of the library-provided resources with the use of scholarly materials accessed from other sources. Faculty members, graduate students, and undergraduate students were studied at several universities. This report focuses on the results from the survey of faculty members from all US universities surveyed. The Lib-Value project is led by a research team at the University of Tennessee, the University of Illinois at Urbana-Champaign (UIUC), Syracuse University, and the Association of Research Libraries (ARL). 6

7 Previous Studies Scholarly reading patterns and information-seeking behavior have been examined through surveys over the past thirty-five years. In 1977 and 1984, national surveys of scientists in the United States were conducted (King et al. 1981). The surveys have been conducted regularly in non-university settings since The first readership survey to be conducted solely in a university setting was completed in 1993 (Belefant-Miller and King 2001). In 2000, the surveys shifted to focus on changing patterns of journal use, due to e-journal publishing, and have been repeated in the U.S., Australia, Japan, and Finland (Tenopir et al. 2010). The surveys found that the increasing prevalence and availability of e-articles encouraged academics to read more articles, though the time spent reading each publication decreased. In the case of United States and Finland, academics who read more articles published more works. However, although Australian academics reported the most e-reading, their reading patterns did not demonstrate a correlation between e-reading amounts and publication. In addition, researchers noted that all academics reported several reasons for readings and varied methods for discovering articles. In 2011, a reading survey was conducted at six United Kingdom higher learning institutions, which includes sections on reading from books and other publications and questions on use of social media (Tenopir et al. 2012). Tenopir and King (2000) and King and Tenopir (2001) summarize reading patterns of science and non-science faculty members through the 1990s. They provide extensive literature reviews and serve as background for the data presented in this report. Other multi-university studies focus on how faculty uses electronic journals, online resources, 7

8 and libraries (Healy et al. 2002). Further studies show that access and convenience, especially electronic access, are important to academic faculty (Maughan 1999). Other studies show the huge impact subject discipline has on reading patterns (Talja and Maula 2003), and different disciplines have varying traditions of the importance of journals compared to other types of information (Fry and Talja 2004). In addition, faculty members in the sciences prefer and read more electronic journal articles than in humanities or social science disciplines (Brown 2003). The results from the U.S. and Australia in 2012 tend to confirm the earlier findings. A 2011 study by the Research Information Network (RIN) found a link between the library and the institution s research performance. Many recent studies have reported on the future of e-books in academia. A report by CIBER (2009) found that nearly two-thirds of teaching staff and students in the United Kingdom have used an e-book to support their work or study or for leisure purposes, and more than half of users said the last e-book they used was provided by their university library. A study at the Health Sciences Library System at Pittsburgh University discovered that over half the surveyed faculty, graduate students, and undergraduate students used library provided e-books for their job duties, and it concluded that respondents are willing to use alternative formats (Folb et al. 2011). Another study at the University of Illinois in 2008 shows that faculty, graduate students, and undergraduate students value the convenience and time saving capabilities this format offers them, as well as the ability to search full-text content of e-books, but there are still disadvantages with its format on the screen (Shelburne 2009). Many other studies have reported similar findings, showing that e-books are becoming a valuable library resource (Chrzastowski 2011; Tenopir et al. 2012). 8

9 A report by CIBER on the use of social media in the research environment found that social media have found applications in the research process, and the most popular tools are those for collaborative authoring, conferencing, and scheduling meetings (Rowlands et al. 2011). The report did not find age to be a good predictor on social media use, but humanists and social scientists used more social media. It concludes social media do not replace traditional material. Methodology Earlier surveys examined just the reading of scholarly articles, but for this survey we expanded it to examine the reading of scholarly books and book chapters and the use and creation of social media. The survey maintained a consistent core of questions and maintained similar questions in each section in order to compare the survey results over time. The questions are based on two principal sections reader-related (demographics) and reading-related. Reader-related questions focus on the demographics of the respondent; the questions include age, gender, percentage of work time spent on various activities, number of personal subscriptions, and two measures of recent academic success publication record and record of recent awards. The reading-related questions mostly use the critical incident technique first developed by Flanagan (1954). The critical incident technique has since been applied to many contexts, including libraries and readings (Radford 2006; Andrews 1991). The survey used the last scholarly reading as the critical incident of reading (Griffiths and King 1991). By asking about a specific most recent reading, respondents should have a better memory of that reading, rather than having to reflect back on multiple readings over a longer period of time. While the last reading may not be representative of a typical 9

10 reading, it allows us to find details and patterns of reading and use. The questions cover many details of that reading, including time spent on the reading, source of reading, purpose of reading, and value of the reading to the purpose. A complete survey instrument is found in the appendix of this report. Starting in February 2012 through October 2012, an message was sent by librarians to approximately 11,332 faculty members at five universities in the United States (Table 1). The message included an embedded link to a survey housed on the University of Tennessee s server. By the last closing date of January 9, 2013, we received 837 responses to the first question for a response rate of 7.4%. 1 Table 1. Response Rates of Participating US Institutions Total Institution Responses Academic Response Rate Staff Seton Hall University % Syracuse University % University of Colorado % University of Illinois, Urbana-Champaign % University of Tennessee, Knoxville % The low response rate may make it hard to generalize across the population, and while our results are not weighted, since we asked demographic questions and know the total population demographic characteristics for some of the questions (age, gender, and discipline, for example), responses and subsequent analysis could be weighted in the future. Weighting the results may help improve the generalizability of the responses. Since respondents were allowed to leave the survey at any time, skip questions, or were timed out automatically if they began the questionnaire and did not complete it, most of the 1 Assumes all invitations were sent to valid and active addresses. 10

11 questions have a lower number of responses. All respondents for a particular question equal 100% for that question. The survey was comprised of four sections: Journal Article Reading, Book Reading, Other Publication Reading, and Demographic Information. 11

12 Demographics of Respondents 12

13 Work Responsibilities Academics at the five US universities spend the most work time on research and writing. Table 2 reflects this, showing faculty members spend just over one third (35%) their time on teaching. Teaching and service (to the department, college, and wider community) also take up large percentages of work time (30% and 17% respectively). Table 2. Percentage of Work Time Spent by US Faculty Respondents Teaching Research & writing Administrative Service Consulting /advising Other Mean Median Mode Percentiles Academic Discipline Twenty-one percent of the respondents are in the social sciences, 18% are in the humanities, and 9% are in the life sciences (Table 3A). We collapsed the disciplines into six categories for analysis (Table 3B) and redistributed the other disciplines into a corresponding category. Psychology, business, and education were combined with social sciences. Computer science, mathematics, and engineering were combined; life and physical sciences were combined, and humanities and fine arts were combined. The remaining other disciplines are disciplines that did not clearly fit into one of the larger categories (i.e., Management/Leadership and Child and Family Studies ). 13

14 Table 3A. Academic Disciplines of US Faculty Respondents Frequency Percent Life sciences Physical sciences Medical science Computer Science Mathematics Engineering Social sciences Business Psychology Education Humanities Fine Arts Law Other Total Table 3B. Academic Disciplines of US Faculty Respondents (Grouped) Frequency Percent Sciences Medical Sciences Engineering/Technology/Math Social Sciences Humanities Others Total Position, Age, and Gender Thirty-six percent of the respondents are professors (Table 4). Other statuses include an archaeological collections specialist, author, consultant, advisor, IT professional, project coordinator, software developer, and visiting scholar. 14

15 Table 4. Academic Position of US Faculty Respondents Frequency Percent Professor Associate Professor Assistant Professor Instructor / Lecturer Adjunct Other (100.0) Academic Staff / (41) (30.8) Professional/Admin Graduate Student (4) (3.0) Research Associate / (49) (36.8) Post Doctoral Other (39) (29.3) Total The respondents ages range from eighteen to one hundred years of age. For analysis we grouped the ages by decade (Table 5). Over a quarter of the respondents are in their thirties, forties, and fifties. Just 18% are over 60 years of age, and 7% in their twenties. Table 5. Age Range of US Faculty Respondents Frequency Percent Under Over Total There are some variations between the respondent s age and their academic position (χ 2 = , p<.0001). Over three-quarters (78%) professors are over fifty years of age, with 41% over sixty years. Sixty-nine percent of associate professors, one quarter of instructors/lecturers, 21% of assistant professors and 18% of adjuncts are in their forty years of age. No professors or associate professors are under 30 years of age. Over half 15

16 (58%) of assistant professors, and just over one-third (36%) of instructors/lecturers are in their thirties. Half (50.3%) of the respondents are male (293 of 582) and half are female (49.7%, 289). The medical sciences, social sciences and humanities are primarily female, while the majority of respondents in the sciences and engineering/technology fields are male (Table 6). Table 6. Gender of US Faculty Respondents by Discipline Male Female Row Total Sciences % % % Medical Sciences % % % Engineering/ Technology/Math % % % Social Sciences % % % Humanities % % % Others % % % Column Total % % % Sixty-nine percent of the professors, 61% of the adjuncts 49% of instructors/lecturers, 42% of associate professors, and 40% of assistant professors are male. Sixty percent of assistant and associate professors, 51% of the instructors/lecturers, and 31% of the professors are female. A majority of respondents over sixty years of age are male (68%). However, women faculty outnumber men in other age groups. Fifty-four percent of faculty in their thirties, 53% in their forties, 52% in their fifties, and 51% under thirty years of age are women. 16

17 Productivity as Measured by Authorship and Awards Authorship has been used as a measure of productivity in past surveys of research universities and in non-university research settings. Over the years it has been shown that faculty who publish more journal articles tend to read more (King et al. 2003). Sixty-nine percent of the respondents published at least one refereed scholarly journal article in the past two years (Table 7). Fewer have published an entire book (15%), but 48% have published a chapter in a book and 30% in a conference proceeding. Taking all these methods of publication together, the average faculty member published six items in the past two years and 78% of the respondents have published at least one scholarly item in the past two years (Table 8). 17

18 Table 7. Number of Items Published in the Last 2 Years by US Faculty Respondents Frequency Percentage Refereed Scholarly Journals ~ ~ > Non-Refereed Journals ~ ~ > Chapters in Books ~ ~ > Conference Proceedings, etc ~ ~ > Entire Books ~ > Table 8. Total Numbers of Publications in the Last 2 Years by US Faculty Respondents Frequency Percent Over Total

19 Discipline significantly influences the number of total publications (F=1.518, p=.182). Respondents in the medical sciences (M=7.06) and engineering/technology/mathematics (M=6.92) published more material in the past two years. Scientists published an average of seven items (M=662) in the past two years, while humanists published five items (M=5.10). Social scientists published the least amount of items (M=4.65). Faculty in engineering/technology disciplines and scientists published more refereed journal articles (Mengineer=4.52, Msciences=4.33), while medical scientists published an average of four articles, social scientists published an average of two articles and humanists published one. All disciplines average less than one entire book publication in the past two years, but social scientists (M=1.28) and humanists (M=1.13) are more likely to publish a book chapter. Older respondents are more likely to have published more in the last two years (p=.048). Respondents over 50 years of age published an average of seven items in the past two years (M=6.50), while respondents under 50 published five items (M=5.02). The differences between number of publications and age may be a result of academic position because the positions, which have older respondents, publish more material (F=6.238, p<.0001). Professors published the most material (M=8.52), followed by associate professors (M=6.17), assistant professors (M=5.97), adjuncts (M=5.58), and instructors/lecturers (M=1.68). Another measure of productivity is whether a respondent has received awards or recognition for their work. We asked respondents whether they received any awards or recognition in the past two years, and then prompted them to describe their award. Thirtyone percent of respondents received an award (180 of 579). The awards and recognitions 19

20 included awards for teaching (e.g., teaching excellence), awards for research (e.g., best paper), awards for dissertations, service awards and other awards. Faculty who won awards also published more material (F=17.451, p<.0001). Award-winning faculty published eight items (M=7.91) and those who did not win an award published five items (M=4.64). Personal Subscriptions We asked how many personal subscriptions to professional journals (in print or electronic form) they receive, including those paid by themselves, received free, or purchased by a grant or other source for personal or shared use. Twenty percent of respondents do not have a personal subscription, and the average number of personal subscriptions is four. Over one-third (37%) of respondents have a print subscription and 54% of respondents have an electronic subscription (Table 9). Forty-eight percent of respondents have a subscription that includes a print and electronic version. The number of personal subscriptions increase with the respondent s age (F=9.141, p<.0001). Respondents over sixty years of age have, by far, more personal subscriptions than younger faculty (M=6.32). Respondents in their fifties have, on average, four personal subscriptions (M=3.89), followed by those in their forties (M=3.38), those in their thirties (M=2.67), and those in their twenties (M=2.41). 20

21 Table 9. Number of Personal Subscriptions for US Faculty Respondents Frequency Percentage Print-only Subscriptions > Electronic-only subscriptions > Print and Electronic Subscriptions > Last Information Source Used While US academics use a variety of sources to inform their work, they may rely on one type of material. We asked, What source did you use for the last substantive piece of information in your work? Journal articles are, by far, the most frequent last source of information (Table 10). A magazine article was the second most frequent source (21%). 21

22 Table 10. Last Information Source Used by US Faculty Respondents Frequency Percent Journal article Book or book chapter Web site Conference proceeding Magazine article Personal contact Other Total Eighty-eight percent of medical science faculty, 80% of science faculty, 66% of social science faculty, and 62% of engineering/technology/mathematics faculty identified journal articles as their last information source used. However, only 35% of humanists used journal articles; instead, the majority of humanists (51%) identified books/book chapters as their last information source used. Seventeen percent of social scientists used book chapters as well. Seventeen percent of engineering/technology/mathematics faculty and 11% of social science faculty used a website. 22

23 Scholarly Journal Article Reading 23

24 Total Amount of Article Reading One of the questions in all of the Tenopir and King surveys from 1977 to the present is an estimate of the total number of articles read in the last month by each respondent. The results provide an approximation of how many articles a respondent reads in a year, which allows us to compare results over time and across populations. Since the question relies on personal recollection, we ask for a relatively short period of time (one month) rather than asking the respondents to reflect back over a longer period of time. We also assume the last month is an accurate representation of a typical month of reading. The first question stated, In the past month (30 days), approximately how many scholarly articles have you read? (Articles can include those found in journal issues, Web sites, or separate copies such as preprints, reprints, and other electronic or paper copies. Reading is defined as going beyond the table of contents, title, and abstract to the body of the article). The actual number is not as important as the relative amounts among types of respondents and over time. For convenience we often report results as readings per year, by taking the monthly number reported by the respondent and multiplying it by 12. In the last month, the faculty read an average of twenty-one articles (M=20.81, SD=21.735). 2 Extrapolated to an entire year, the average US faculty member reads 252 articles. Only 6% of the respondents report zero readings in the past month; zero readings is included in our average. 2 Excludes outliers over 150. Including outliers the mean is

25 Last Incident of Reading and Date of Publication The next set of questions asks the respondents to focus on the last scholarly article they read. This variation of the critical incident technique assumes the last article reading is random and provides detailed information on a random sample of the readings by faculty members. We asked, The following questions in this section refer to the SCHOLARLY ARTICLE YOU READ MOST RECENTLY, even if you had read the article previously. Note that this last reading may not be typical, but will help us establish the range of reading patterns. We then asked for the title or topic of the journal article from which the last reading took place in order to focus their minds on the article for the rest of the critical incident questions. The next question asked for the publication or posting date of the last article reading. In the surveys in the U.S. from 1977 to 2005, we have seen an increase in reading of articles older than the first year of publication, though reading is still skewed to the most recent articles (King et al. 2009). In the surveys in the U.S. and Australia in 2005, we found an increase in the reading of older articles, with just half of readings within the first year of publication, and in the U.K. in 2011 nearly half of the readings are from articles in their first eighteen months of publication (Tenopir et al. 2012). This differs from older studies, which found about two-thirds of reading within the first year of publication (Tenopir et al. 2005). The change may be a result of availability of electronic back files, an increase in the respondent s searching capabilities to identify older articles, and/or search system features such as relevance ranking that allows older articles to be more accessible. There are, of course, some differences based on subject discipline, with medical staff reading a higher proportion of current articles. 25

26 Thirty-nine percent of article readings by faculty members are within the first year of publication (Table 11). The year of publication ranges from as early as 1944, with 10% published before Table 11. Age of Article Reading by US Faculty Year Frequency Percentage Over 15 years (Before 1997) ~ 15 years ( ) ~ 10 years ( ) ~ 5 years ( ) Year (2011) Less than 1 year (2012) Total Graduate students and undergraduate students report more readings over two years old than do faculty members. Thirty-one percent of the readings by undergraduate students (155 of 428) and 21% of the readings by graduate students (202 of 958) in the United States are in their first year of publication, while 39% of the readings by faculty members are in the first year of publication. Studies done by Guthrie (2000), Odlyzko (2000), and Herman (2004) provide further research on the life of a journal article and its half-life. They found many older articles are heavily used when they are conveniently accessible; however, academics tend to cite more recent articles in order to seem current and up-to-date in their field. Their research further suggests that back files are a key investment in addition to current subscriptions. 26

27 Novelty of Information in the Reading Since this is a random sample of article readings rather than unique articles, the article may have been previously read. In this study, 22% of the article readings are rereadings. We also wanted to find out the reader s knowledge of the article content before this reading (i.e., was the information familiar to them before the reading). Together, these questions indicate if articles are often used as sources of new information. Over two-thirds (70%) of the respondents say they knew parts of the information in the article prior to this reading, but only 5% knew all (or a majority) of the information. To further determine the novelty and value of articles as sources of new information, we asked those who knew about all or part of the information in the article reading where they originally found it. Another journal article, informal discussion with colleagues, and other sources are the main sources of information found in articles (Table 12). The other responses include a database search, an author, a clinical case, books, graduate school, prior research, a student, and notices. Table 12. Source of Information Not Obtained Through Last Article Reading by US Faculty Frequency Percent Conference or workshop Informal discussion with colleagues Listserv or blog Journal article from colleague Preprint/e-print service (e.g., arxiv.org) Web site of author Institutional Repository Other Total